India's AI-DPI Integration: A New Frontier for State Capacity and Governance
India's digital public infrastructure (DPI) — including Aadhaar, UPI, and GST — has already transformed governance by reducing fraud, improving efficiency, and catalysing digital payments. The next leap, experts say, lies in integrating artificial intelligence (AI) with this infrastructure to further enhance state capacity and citizen services.
Speaking at a recent conference at the National Council of Applied Economic Research (NCAER), analysts highlighted three reasons for optimism: India's status as the world's largest natural data laboratory, the potential for AI to reimagine the Indian state, and the possibility that AI deployment may create more jobs than AI development.
India has rapidly transitioned from being data poor to data rich, thanks to the digital exhaust from DPI systems. UPI generates 23 billion transactions monthly, GST processes 20 crore payments, and Aadhaar records 27 billion authentications annually. This data offers unprecedented insights into financial inclusion, welfare targeting, and economic activity — without relying on delayed surveys.
However, challenges remain. Digital silos, limited internal technology capacity within government, and a focus on PDFs rather than APIs have hindered full interoperability. AI's ability to process unstructured data could help overcome these barriers, enabling a more seamless citizen interface.
Moreover, AI-powered language interfaces could open up labour and education markets, while improving public health expenditure. India could become the first large country to integrate AI into population-scale DPI, provided it invests in rigorous research, think tanks, and open public data commons.
While weak state capacity has historically hampered India's economic transformation, DPI combined with AI offers a path to accelerate development and improve governance without massive administrative expansion. The key lies in leveraging data responsibly and ensuring interoperability across systems.
